Content-oriented learned image compression

TL;DR

This paper introduces a content-oriented learned image compression approach that considers image semantics, achieving competitive subjective quality compared to existing methods by handling different content types with tailored strategies.

Contribution

It proposes a novel content-aware compression method that incorporates image semantics into the optimization process, unlike previous models that ignore content.

Findings

Achieves competitive subjective image quality.

Handles various image contents with specialized strategies.

Outperforms classic methods in perceptual quality.

Abstract

In recent years, with the development of deep neural networks, end-to-end optimized image compression has made significant progress and exceeded the classic methods in terms of rate-distortion performance. However, most learning-based image compression methods are unlabeled and do not consider image semantics or content when optimizing the model. In fact, human eyes have different sensitivities to different content, so the image content also needs to be considered. In this paper, we propose a content-oriented image compression method, which handles different kinds of image contents with different strategies. Extensive experiments show that the proposed method achieves competitive subjective results compared with state-of-the-art end-to-end learned image compression methods or classic methods.